FRENCH MBTI STEP I DATA SUPPLEMENT July 2005

Transcription

1 FRECH MBTI STEP I DATA SUPPLEMET July 2005 John Hackston, OPP Ltd 2005 OPP Ltd. All rights reserved. Myers-Briggs Type Indicator and MBTI are registered trade marks of the Myers- Briggs Type Indicator Trust. OPP Ltd is licensed to use the trade marks in Europe. OPP and the OPP logo are registered trade marks of OPP Ltd.

2 TABLE OF COTETS Introduction...3 Type Distribution...4 Business Studies Students...4 Training Course Delegates...6 OPPassessment Data...8 Type Table Comparison...9 Internal Consistency Reliability...10 Validity; the Accuracy of the French MBTI Step I Instrument in Predicting Best-Fit Type...11 Group Differences in Type...13 Gender...13 Age...14 Job Role and Occupational Level...14 Education...16 Work Area...16 ationality...18 Employment Status...18 References...19 Appendix 1: Sample Descriptions...20 Sample 1: Business Studies Students...20 Sample 2: Delegates on MBTI Qualifying Training...20 Sample 3: Data from OPPassessment

3 ITRODUCTIO Personality questionnaires have a special place in France. The Myers-Briggs Type Indicator (MBTI ) has received particular attention, in part because of its special character as a type indicator. This theoretical base and the research studies which result from this are different from those of trait questionnaires; however the psychometric qualities of the MBTI instrument are well-established and have been continually reinforced by new research. The reliability and validity of the instrument are very good, and France, where the questionnaire is widely used, is no exception to this rule. The results of the research here demonstrate the commitment of the publishers, distributors and users of the MBTI instrument to continued research and development. This update draws upon data collected from a number of samples between 2002 and 2005: 363 business studies students from two centres (Lille and ice) who completed the MBTI instrument for research purposes and as part of their own development 612 delegates on MBTI qualifying training programmes held in France from January 2002 to March individuals who completed the MBTI Step I or Step II 1 instrument in French via the OPPassessment system 2. Data from training programmes has been excluded from this group, and the sample is therefore likely to be a representative sample of the groups of people with whom the French MBTI instrument has and will be used for development, counselling, team-building, etc. As such, it is likely to represent a cross-section of the Francophone European professional and managerial population. A fuller description of each sample is given in Appendix 1. This data supplement contains a number of research studies based on this data, including: Type distributions; type tables showing the proportion of each type within the three samples. The three French groups are similar to other comparable groups across the world. Internal consistency reliability. All four dimensions show good reliability, above 0.7, in all groups. Validity. There is a good match between the results of the questionnaire and best-fit (validated) type. Respondents are confident about their results, and there is no evidence that a new and specifically French scoring algorithm would improve the accuracy of the instrument. Group differences in Type. Gender, age, job role, occupational level, education and employment status are examined. 1 Step II is a longer version of the MBTI instrument which splits each of the four MBTI dichotomies into 5 subscales or facets. The Step II instrument does however also yield the standard Step I scores and 4-letter type, and these have been used in the analyses described in this document. 2 OPPassessment allows personality questionnaires such as the MBTI instrument to be administered via

4 TYPE DISTRIBUTIO Type tables are a way of illustrating the proportion of each type within a particular group. For each of the 16 different types, the number of cases, the percentage of the total that this represents, and the Self Selection Ratio (SSR) is shown. The SSR (Myers et al 1998) is a way of demonstrating whether a given type appears more or less often in a particular group than would be expected compared to a reference group. An SSR greater than 1 indicates that a type is over-represented, and an SSR of less than 1 that it is under-represented. Here, the SSR has been calculated in comparison to the UK general population (Kendall, 1998). The UK general population has been chosen as a useful general population reference group as a large representative French group does not currently exist. Other evidence (e.g. Hackston and Kendall, 2004; Quenk et al, 2004) does however suggest that although behaviour varies greatly from culture to culture, the frequencies of underlying psychological types do not. In this section of the update, type tables for each of the three samples are presented. BUSIESS STUDIES STUDETS Table 1: Type Tables for Business Studies Students Reported Type (=363) ISTJ ISFJ IFJ ITJ Type n % n = % SSR=0.46 n = % SSR=0.28 n = % SSR=3.41 n = % SSR=3.71 ISTP ISFP IFP ITP n = % SSR=0.69 n = % SSR=0.54 n = % SSR=2.16 n = % SSR=4.38 ESTP ESFP EFP ETP n = % SSR=0.62 n = % SSR=0.38 n = % SSR=2.27. n = % SSR=2.18 ESTJ ESFJ EFJ ETJ n = % SSR=0.61 n = % SSR=0.55 n = % SSR=2.86 n = % SSR=1.90 E I S T F J P % 46.0% 37.7% 62.3% 47.9% 52.1% 47.7% 52.3% - 4-

5 Best Fit Type (=363) ISTJ ISFJ IFJ ITJ Type n % n = % SSR=0.56 n = % SSR=0.41 n = % SSR=3.24 n = % SSR=4.93 ISTP ISFP IFP ITP n = % SSR=0.64 n = % SSR=0.46 n = % SSR=3.00 n = % SSR=3.46 ESTP ESFP EFP ETP n = % SSR=0.57 n = % SSR=0.41 n = % SSR=2.49. n = % SSR=2.18 ESTJ ESFJ EFJ ETJ n = % SSR=0.56 n = % SSR=0.50 n = % SSR=2.18 n = % SSR=1.03 E I S T F J P % 50.1% 38.8% 61.2% 45.2% 54.8% 46.6% 53.4% For both reported and best-fit type, there is a clear majority of people with preferences for Intuition; EFP is the most common whole type preference in both cases. Compared to the general population reference group, those with a preference for Intuition are over-represented. This is typical of student groups and of those who have been educated to a higher level (for example, see MacDaid et al, 1991). - 5-

6 TRAIIG COURSE DELEGATES Reported type results from the MBTI instrument were available for almost the entire group, and best-fit (validated) type was available for almost all of these. Table 2: Type Tables for Training Course Delegates Reported Type (=597) ISTJ ISFJ IFJ ITJ Type n % n = % SSR=0.36 n = % SSR=0.29 n = % SSR=3.53 n = % SSR=4.21 ISTP ISFP IFP ITP n = 4 0.7% SSR=0.11 n = 8 1.3% SSR=0.21 n = % SSR=3.97 n = % SSR=2.58 ESTP ESFP EFP ETP n = % SSR=0.64 n = % SSR=0.51 n = % SSR=2.83 n = % SSR=2.11 ESTJ ESFJ EFJ ETJ n = % SSR=0.57 n = % SSR=0.53 Best-Fit Type (=578) n = % SSR=3.21 n = % SSR=1.79 E I S T F J P % 41.5% 31.3% 68.7% 38.4% 61.6% 47.4% 52.6% ISTJ ISFJ IFJ ITJ Type n % n = % SSR=0.26 n = % SSR=0.35 n = % SSR=4.06 n = % SSR=4.36 ISTP ISFP IFP ITP n = 7 1.2% SSR=0.19 n = % SSR=0.43 n = % SSR=4.75 n = % SSR=2.54 ESTP ESFP EFP ETP n = % SSR=0.48 n = % SSR=0.52 n = % SSR=3.24 n = % SSR=2.39 ESTJ ESFJ EFJ ETJ n = % SSR=0.27 n = % SSR=0.47 n = % SSR=2.71 n = % SSR=1.14 E I S T F J P % 46.0% 27.7% 72.3% 32.4% 67.6% 40.5% 59.5% The most frequent type preferences are EFP and IFP; overall the group tends to have preferences for Intuition and Feeling, and to a lesser extent for Extraversion - 6-

7 and Perceiving. Most of these results are slightly more pronounced still when best-fit type is used. Compared to the general population, those with a preference for Intuition are particularly over-represented. Although not typical of the general population, similar results (especially with regard to Intuition) have been found with other groups of MBTI users and training course delegates. The table below contrasts a number of these groups from around the world with the two large general population groups which currently exist for the MBTI. These data relate to reported type. Table 3: Comparison of French Workshop Participants with other Groups Percentage of Each Type E I S T F J P US nationally representative sample UK nationally representative sample French MBTI workshop participants Australian MBTI workshop participants British MBTI workshop participants Canadian MBTI workshop participants Dutch MBTI users German MBTI workshop participants Indian MBTI workshop participants Japanese MBTI users Median of MBTI users Australian, Canadian, Indian and Japanese compiled for an International panel at the 2004 Association of Psychological Type International conference. Other groups from OPP data. - 7-

8 OPPASSESSMET DATA Table 4: Type Table for OPPassessment Data (Reported Type, =916) ISTJ ISFJ IFJ ITJ Type n % n = % SSR=0.88 n = % SSR=0.35 n = % SSR=1.12 n = % SSR=3.43 ISTP ISFP IFP ITP n = % SSR=0.61 n = % SSR=0.20 n = % SSR=1.56 n = % SSR=2.50 ESTP ESFP EFP ETP n = % SSR=0.71 n = % SSR=0.39 n = % SSR=0.98 n = % SSR=2.46 ESTJ ESFJ EFJ ETJ n = % SSR=1.67 n = % SSR=0.56 n = % SSR=1.64 n = % SSR=3.79 E I S T F J P % 39.4% 53.6% 46.4% 66.2% 33.8% 63.3% 36.7% The most common single type preference is ESTJ (17% of the total); this is a common finding with managerial groups. The SSR results suggest that, in comparison to the general population, those with preferences for T are overrepresented, and those with preferences for SF are under-represented. Again, this is a common finding with managerial groups. - 8-

9 TYPE TABLE COMPARISO As a further comparison, the percentage of people of for each type dichotomy (E compared with I, S with etc.) for the three groups are compared with a number of other reference groups in the table below: Table 5: Comparison of the Three Groups with other Relevant Groups Group Percentage of Each Type E I S T F J P French business studies students French MBTI workshop participants French OPPassessment group Median of MBTI Users French Ashridge delegates Total Ashridge delegates US leadership development group US nationally representative sample UK nationally representative sample The French groups are more similar to other relevant comparison groups than to each other. For example, the OPPassessment group is similar to Ashridge delegates and the leadership development sample. 4 See Table 3 above 5 French delegates to management development programmes at Ashridge Management School, UK. =263. See Carr et al, Total Ashridge group (n=8,039), of which the French delegates are a subset; contains data from managers from 86 different countries, 96% of them European. 7 =26,477. See Fleenor, =3,009. See Myers et al, =1,634. See Kendall,

10 ITERAL COSISTECY RELIABILITY The reliability of a test or questionnaire relates to how consistent and precise it is. The internal consistency reliability addresses the question of whether all the questions in a scale measure the same construct. A common measure of internal consistency reliability is coefficient alpha (an average of all question to question correlations). The alpha coefficients for the French samples are shown in Table 6 below: Table 6: Internal Consistency Reliability Dimension Business Students Coefficient Alpha Training Delegates OPPassess. (Step I) OPPassess (Step II) 10 E-I S T-F J-P It is generally agreed that internal consistency reliability should achieve a value of at least 0.7. On this basis, all the dimensions of the questionnaire show good reliability in all groups. 10 Estimated by scoring the Step II data using the Step I scoring key - 10-

11 VALIDITY; THE ACCURACY OF THE FRECH MBTI STEP I ISTRUMET I PREDICTIG BEST-FIT TYPE The purpose of the MBTI instrument is to help individuals to establish their true, validated or best fit psychological type. A key measure of the validity of the instrument is, therefore, how well the results relate to best-fit (validated) type. Best-fit data was available for two of the samples. The business studies students were given group feedback on their results, and their best-fit type was collected alongside their reported type results. The training delegates established their best fit type as part of their training programme, and this was collected for almost the entire sample (578 people). Table 7 below presents these results alongside equivalent data from another (UKbased) best-fit study using the Step I questionnaire (Kendall, 1998). The French questionnaire performs in a very similar way to the English version, and there is very good evidence for the accuracy of the instrument. Table 7: Match of Reported and Best-Fit Type Training Delegates (=578) Business Studies (=363) UK Study (=386) Agrees with all 4 letters 67.8% 62.3% 71.5% 93.1% 93.4% Agrees with 3 letters 25.3% 31.1% 21.5% 93.0% Agrees with 2 letters 6.1% 4.9% 6.1% Agrees with 1 letter 0.7% 6.9% 1.7% 6.6% 0.3% 7.0% Agrees with no letters 0.2% 0.0% 0.3% Percentage Agreement Type Dichotomy Training Business UK Study Delegates Studies E-I 90.1% 89.8% 92.1% S- 91.0% 91.6% 93.8% T-F 88.3% 86.8% 88.4% J-P 90.6% 86.8% 89.0% Two further sets of analysis were carried out to investigate the validity and accuracy of the questionnaire. Training course delegates were asked how confident they felt about their results on each type dichotomy (on a scale from 1 to 5, where 5 indicated the highest degree of confidence). For every dimension, over 80% of the group were confident about their type, providing further support for the validity of the MBTI approach. Detailed results are shown in Table 8 below: Table 8: Degree of Confidence in Results Degree of Percent of Group Confidence E-I S- T-F J-P 5 65% 61% 58% 63% 4 19% 23% 23% 22% 3 11% 10% 14% 10% 2 5% 4% 3% 3% 1 1% 2% 2% 2% % at 4 or above 84% 84% 81% 85% - 11-

12 Secondly, item-level data from the business students sample was used to recalculate prediction ratios for each item. From these prediction ratios revised item weightings were derived 11, and these were then applied to the data to produce revised reported types for each person. These were then compared with best-fit type. The results showed no improvement over the level of agreement achieved using the existing Step I item weightings, as well as a high level of agreement between the new weightings and the standard Step I item weightings. There was therefore no evidence to suggest that a different scoring system should be applied to the MBTI Step I instrument in France. In summary, then, there is good evidence for the validity of the French MBTI Step I instrument. Specifically: There is a high level of agreement between best-fit and reported type, as high as for the English language version Respondents are overwhelmingly confident about their results There is no evidence that a specifically French scoring algorithm would improve the accuracy of the instrument. 11 For a description of how prediction ratios are derived and then used to devise scoring weights, see Myers and McCaulley (1985), pages

13 GROUP DIFFERECES I TYPE Across the three samples, a variety of different demographic information was collected, as shown in table 9 below: Table 9: Available Demographic Information Business Students Training Delegates OPPassessment Gender Age Job Role Occupational Level Education Work Area ationality Employment Status The relationship of Type to each of these factors is described below. GEDER Most groups who take the MBTI instrument show a significant gender difference on the Thinking-Feeling dimension, and this is the case for the three groups in this study, as shown in Table 10 below: Table 10: Gender Differences on the T-F Dimension Group Gender T F Business Students (Χ 2 = 12.96, df=1, Male % within gender 12 % within Type % 62.8% sig at 0.1% level) Female % within gender % within Type Training Delegates (Χ 2 = 6.76, df=1, Male % within gender % within Type sig at 1% level) Female % within gender % within Type OPPassessment (Χ 2 = 62.48, df=1, sig at the 0.1% level) Male Female % within gender % within Type % within gender % within Type % 37.2% % 39.1% % 60.9% % 73.6% % 26.4% % 43.9% % 56.1% % 28.3% % 71.7% % 47.2% % 52.8% 12 For example, 63.5 male business students have a T preference and 36.5% have an F preference 13 For example, 62.8% of T business students are male, and 37.2% are female

14 Thinking preferences are over-represented amongst men and Feeling preferences are over-represented amongst women. This effect has been found many times with many different versions of the instrument in a number of different cultures. Here, the difference is particularly clear amongst those with a dominant Thinking or Feeling and an auxiliary Sensing function. AGE one of the samples showed a statistically significant relationship between Type and age. JOB ROLE AD OCCUPATIOAL LEVEL Previous research has demonstrated that those in higher level jobs are more likely to have preferences for Intuition and for Thinking than those in lower level jobs. This is reflected in the relationship of the Thinking-Feeling dimension with job role (function) in the training sample and of both Sensing-Intuition and Thinking-Feeling with occupational level in the OPPassessment sample. Table 11: Thinking-Feeling and Job Role 14 Fonction T F PDG/Direction Générale % within Fonction % % Cadre Supérieur % within Fonction % % Cadre % within Fonction % % Technicien/Agent de Maîtrise % within Fonction % % Employé % within Fonction 0 0.0% % Profession libérale/consultant % within Fonction % % (ote: only those roles with a total sample of 10 or more are shown) Managers and executives are much more likely to have preferences for Thinking than are other groups. A similar pattern is found in the data from OPPassessment; those with preferences for Intuition and Thinking are over-represented at a higher level, as shown in tables 12 and 13 below. 14 Χ 2 = 24.39, df=10, sig at 1% level

15 Table 12: Sensing-Intuition 15 and Occupational Level Occupational Level S Top level % % Senior executive % % Upper middle management % % Middle management % % First level management/supervisor % % Employee % % Other % % (Total) % % % Table 13: Thinking-Feeling 16 and Occupational Level Occupational Level T F Top level % % Senior executive % % Upper middle management % % Middle management % % First level management/supervisor % % Employee % % Other % % (Total) % % % ote also that in this data set as a whole, preferences for Intuition and Thinking are over-represented in comparison to the general population. 15 Χ 2 = 15.14, sig at the 5% level. 16 Χ 2 = 22.72, sig at the 0.1% level

16 EDUCATIO Analysis by education was not carried out for the business student sample. The training delegate data showed a small but statistically significant tendency 17 for those at an educational level of Bac +5 and above to be more likely to have a Perceiving preference than those at the Bac + 3 & 4 level. Specific educational qualifications were not available for the OPPassessment sample; however the age at which individuals left full-time education was. Those who left full-time education at a greater age were significantly more likely to have preferences for Intuition 18 and for Perceiving 19. WORK AREA Previous type research would suggest that an individual s type would influence their choice of career, and indeed there is a statistically significant relationship between S-, T-F, J-P and job type. In the tables below, categories have been re-ordered according to the percentage of E, S, T or J, and job types with less than 20 respondents have been omitted. Table 14: Sensing-Intuition 20 and Work Area Job Type S Administrative or secretarial % % Finance % % Sales, customer service % % Science, engineering % % Other private sector % % Other % % IT % % HR, training, guidance % % Other business services % % Research and development % % (Total) % % % 17 Χ 2 = 11.23, df=5, sig at 5% level. 18 Independent-samples t-test; t=-5.198, sig at the 0.1% level 19 Independent-samples t-test; t=-2.735, sig at the 1% level 20 Χ 2 = 25.21, sig at the 5% level

17 Table 15: Thinking-Feeling 21 and Work Area Job Type T F Science, engineering % % IT % % Finance % % Research and development % % Sales, customer service % % Other private sector % % Other % % Other business services % % HR, training, guidance % % Administrative or secretarial % % (Total) % % % 21 Χ 2 = 45.27, sig at the 0.1% level

18 Table 16: Judging-Perceiving 22 and Work Area Job Type J P Administrative or secretarial % % Research and development % % Finance % % Science, engineering % % IT % % Sales, customer service % % Other % % Other private sector % % Other business services % % HR, training, guidance % % (Total) % % % ATIOALITY For the OPPassessment group, information on nationality was available. Although half the group were French, other nationalities (e.g. Belgian, Swiss) were also represented (see Appendix 1 for details). Analysis suggested that the Belgian subgroup was significantly more likely 23 to have a Sensing preference than the French. This sample should however be treated with caution, as 55% of the Belgian sample came from one organisation. EMPLOYMET STATUS Employment status (available for the OPPassessment sample) showed a relationship with the Thinking-Feeling dimension; those who worked part-time were more likely than other groups to have preferences for Feeling. This is likely to be a gender effect; 83% of part-time workers were female, compared with 35% of the total group and 33% of full-time workers. 22 Χ 2 = 30.72, sig at the 1% level. 23 Χ 2 = 11.77, sig at the 0.1% level

19 REFERECES Carr, M., Curd, J. and Dent, F. (2004). Ashridge Management School MBTI Research into Distribution of Type. Ashridge Management School. Cauvin, P et Cailloux, G (1999) Données statistiques in Manuel de formation qualifiante au MBTI Présentées à la conférence internationale de Scottsdale 1999 Cauvin, P et Cailloux, G (2005) L'instrument MBTI in Les types de personnalité ESF Paris 6 édition 2005 Fleenor, J. W. (1997). The Relationship Between the MBTI and Measures of Personality and Performance in Management Groups. In Developing Leaders. Davies-Black. Palo Alto, CA. Hackston, J. and Kendall, E. (2004). Step II and Culture. Paper presented at the APT international conference, Toronto, July Kendall, E. (1998). Myers-Briggs Type Indicator European English Edition Step I Manual Supplement. CPP, Inc. Palo Alto, CA. MacDaid, G. P., McCaulley, M. H. and Kainz, R.L. (1991). Atlas of Type Tables. Center for Applications of Psychological Type, Gainesville, FL. Myers, I. B. and McCaulley, M. H. (1985). Manual: A Guide to the Development and Use of the Myers-Briggs Type Indicator. CPP Inc.; Palo Alto, CA. Myers, I. B., McCaulley, M.H., Quenk,.L. and Hammer, A. L. (1998). MBTI Manual: A Guide to the Development and Use of the Myers-Briggs Type Indicator. CPP, Inc. Palo Alto, CA. Quenk,. L., Hammer, A. L. and Majors, M. (2004). MBTI Step II Manual European Edition. CPP Inc. Palo Alto, CA. <OTE: it would be useful in the text to have some specifically French references and to add them in here> - 19-

20 APPEDIX 1: SAMPLE DESCRIPTIOS SAMPLE 1: BUSIESS STUDIES STUDETS The sample consisted of 363 business studies students. 197 (54%) were female and 166 (46%) were male. Age ranged from 18 to 22 years with a mean of 20 years. SAMPLE 2: DELEGATES O MBTI QUALIFYIG TRAIIG The sample consisted of 612 delegates to French MBTI training programmes from January 2002 to March (69%) were female and 189 (31%) male; age ranged from 24 to 63 (with an average age of 42 years). The majority (563 people, 96% of those who answered the question) said that overall they were satisfied with their job. Most of the group (570 people, 93%) gave information as to their job role; the majority of these saw themselves as professionals/consultants (41%) or as managers or executives (49%): Fonction % PDG/Direction Générale Cadre Supérieur Cadre Technicien/Agent de Maîtrise Employé Ouvrier 0 0 Profession libérale/consultant Etudiant Retraité/Bénévole Demandeur d emploi Artisan/Commerçant Agriculteur 0 0 Sans profession Just under three-quarters of the group (445 people, 73%) also stated their area of work. ot surprisingly, most of these were in training and development (66%) or HR (21%): Service % Direction Générale Production Achat Marketing/Commercial Etudes/R&D Administration/Juridique Finances Compatabilité Informatique Gestion du personnel Formation/Développement

21 Almost all the group (604 people, 99%) gave their educational background: iveau d études % CAP/BEP < Bac Bac Bac + 1 & Bac + 3 & Bac + 5 et au- delà SAMPLE 3: DATA FROM OPPASSESSMET This sample consists of 916 individuals who completed the MBTI instrument in French via the OPPassessment system. Results from over 90 different organisations are included. 65% of the respondents were male, and 35% were female. Age ranged from 21 to 65 years, with a mean of 38 and a median of % of respondents stated their nationality; while half were French, other nationalities were also represented: ationality Percentage French 49.5% Belgian 31.6% Swiss 7.8% Other European 7.7% Other 3.4% The majority of the group were in full-time employment: Employment Status Percentage Full-time 90.7% Part-time 4.9% Self-employed 2.9% Unemployed 1.2% Retired 0.1% Homemaker 0.1% The majority of the group were of managerial level or above: Occupational Level Percentage Top level 2.2% Senior executive 31.0% Upper middle management 21.3% Middle management 16.3% First level management/supervisor 5.1% Employee 19.9% Other 4.1% - 21-

22 A range of work areas were represented: Work Area (Job Type) Percentage Finance 16.6% HR, training, guidance 15.3% Sales, customer service 13.7% IT 11.3% Science, engineering 7.3% Business services 6.3% Research and development 5.9% Admin or secretarial 3.9% Health, social services etc. 0.4% Land, sea or air transport 0.4% Leisure, personal service 0.3% Other private sector 6.4% Other public sector 1.0% Other 11.2% - 22-